Decisions about whether to be tested for genetic risk of breast cancer are difficult. There are qualitative and quantitative dimensions of this decision. Quantitative dimensions include understanding conditional probabilities, relative and absolute risk, and the logic of statistical risk models. Qualitative dimensions include understanding what is breast cancer, what does genetic risk for breast cancer mean, what people should do in the event of positive and negative test results, and deciding under what circumstances a person should consider being tested. Aims. The goals of this project are to understand how women who have never had cancer themselves decide about whether to undergo predictive testing for genetic risk of breast cancer, and to develop and test a web-based computerized Intelligent Tutoring System (ITS) to help women make this decision using information already vetted, approved, and available on the National Cancer Institute web site. The first aim is better understand decision-making processes. The second aim is to develop a web- based AutoTutor, a sophisticated ITS with an animated conversational agent. Innovation. This is, we believe, the first use of an ITS to improve patients' medical decision making. These tutorials will teach women about the qualitative and quantitative concepts related to predictive testing. The ultimate goal is helping women make better decisions about genetic testing for breast cancer risk. Methods. Dimensions of this research and development project are developing the web-based AutoTutor; conducting randomized controlled experiments; and carrying out fine-grained cognitive analyses. The fine-grained analysis will integrate detailed process data with outcomes and posttest responses from 120 participants. The AutoTutor will be developed and tested in three phases corresponding to two tutor modules emphasizing qualitative and quantitative content, and a post-production phase. This will be accomplished through an iterative process with cycles of (1) preliminary research, (2) tutor development, (3) empirical research, and (4) tutor revision. New dependent measurers will be developed in a study with 60 participants. Three controlled experiments will empirically test the AutoTutor and assess decision-making. Two experiments of 120 participants each will address each module and a third web-based experiment with 80 participants will test the complete tutor. Participants will be randomly assigned to the AutoTutor, the National Cancer Institute web site or a control group receiving unrelated information. We will work from the beginning to lay the foundations for the next, more sophisticated generation of the AutoTutor. Personnel. PIs Christopher Wolfe at Miami University and Valerie Reyna at Cornell University have considerable experience with research on medical decision-making, learning technologies and web-based interventions, web-based psychology experiments, quantitative decision making, and verbal reasoning. Expert consultants are Nananda Col MD, breast cancer expert and director of the Center for Outcomes Research and Evaluation, Maine Medical Center, and genetic counselor Sara Knapke. PUBLIC HEALTH RELEVANCE: The goal of this project is to develop a web-based Intelligent Tutor about qualitative and quantitative dimensions of the decision to undergo predictive testing for genetic risk of breast cancer. The purpose is to understand how women make this decision and help improve decision making. Research methods include randomized controlled experiments and fine-grained cognitive analysis.